The short version
If you sell work over the phone, the biggest hole in your month is almost never the website. It is the calls that rang out. A contractor doing forty inbound calls a week and missing eight of them is throwing away roughly a fifth of everything the marketing paid for, and no SEO report will ever show it, because the report stops counting at the moment the call connects. You can measure this today. Pull the call log from your phone provider, filter for inbound calls under fifteen seconds and calls with no answer, and put them on a chart by hour of day. Most shops find a wall of them between three and six in the afternoon, which is exactly when a crew is packing up and the office has gone quiet. Fixing that hour is cheaper and faster than any ranking project you could start this quarter.
Where the money actually goes missing
Everybody in this business argues about rankings. Almost nobody counts how many times the phone rang and nobody answered it, which is strange, because for a contractor that is usually the single largest loss in the entire chain.
Think about what the marketing is actually buying. Somebody has a problem, they search, they land on a page, they read enough to trust you, and they tap the number. Every dollar of that was spent to produce one event: a phone ringing in your office. If it rings out, all of it was spent for nothing, and the worst part is that the spend looks fine on paper. Impressions went up. Clicks went up. Calls from the site went up. Revenue did not.
We started asking for call logs during audits because the numbers stopped adding up on their own. A shop would show us a good traffic chart and a bad month, and there was no explanation anywhere in the site data. The explanation was always in the phone data, and the phone data belonged to a different vendor nobody had ever looked at.
What the logs keep showing us
The pattern is consistent enough that we now predict it before we look. Three clusters, every time.
The first is the afternoon wall. Between roughly three and six, the office empties out, the crews are wrapping up, and whoever was covering the phone is now doing something else. That window is also when homeowners get off work and start making the calls they have been putting off all day. The busiest inbound hour of the day lines up almost perfectly with the hour you are least able to answer, and that overlap is where most of the loss lives.
The second is the sub-fifteen-second call. It connects, it rings, the person hangs up. Most providers log this as a completed call, so it shows up in your call count as a success. It was not a success. Someone decided in twelve seconds that you were not going to answer and went to the next result, and there is a decent chance they were already on hold with your competitor before your voicemail finished its greeting.
The third is the lunch gap. Shorter than the afternoon wall and easier to fix, because it is one person taking a break rather than a structural staffing problem.
Voicemail is not a safety net
The standard response to all of this is that the caller can leave a message. In practice, very few do, and the ones who do are not evenly distributed. The person calling about a scheduled maintenance visit will leave a message. The person standing in a flooded basement will not. Voicemail filters out exactly the urgent, high-value, ready-to-buy calls you most wanted, and leaves you the ones that could have been an email.
Then there is what happens after. A voicemail that gets returned the next morning is competing against a company that picked up yesterday afternoon and is already on site. You are not calling a lead back. You are calling a stranger who solved their problem eighteen hours ago and now has to tell you no.
How to put a dollar figure on it
This calculation takes about twenty minutes and it tends to end arguments.
Count your missed and sub-fifteen-second inbound calls for one full month. Multiply by the share of answered calls that turn into booked jobs, which you probably already know within a few points. Multiply that by your average job value. That is the low estimate of what the month cost you, low because it assumes a missed caller converts at the same rate as an answered one, and missed callers skew urgent, which means they skew toward the jobs you actually want.
For a shop with a four thousand dollar average job and a one in three close rate, thirty missed calls a month is forty thousand dollars of work that walked. Nobody is spending forty thousand a month on marketing. The leak is bigger than the budget.
What to change, in the order worth changing it
Start with coverage of the worst hour. Not all hours, just the one the chart says is bleeding. That might be a single person whose only job from three to six is the phone, or it might be an answering service that takes the details and books, rather than one that reads a script and promises a callback. The test for whether an answering service is worth paying for is simple: can it put a job on your calendar? If it can only take a message, you have bought a more expensive voicemail.
Second, set a callback rule and give it to one named person. Every missed call gets a callback attempt within ten minutes during business hours, and the log gets cleared before anyone goes home. Ten minutes is not arbitrary. The gap between calling back in ten minutes and calling back in two hours is the difference between catching someone who is still looking and catching someone who already hired somebody.
Third, add a text fallback on the same number. A caller who rings out gets an automatic message saying you saw the call and asking what the job is. A decent share of people will answer that text who would never have left a voicemail, and the ones who do are handing you their number and their problem in writing.
Fourth, and only fourth, look at the website. The order matters. Fixing the site so it produces thirty percent more calls, while a fifth of calls ring out, means you have paid to increase the number of people who hear your voicemail.
Where the phone meets the ranking
There is a search angle here, and it is not the one people expect. Your Google Business Profile is a call button, and when a caller taps it and nobody answers, the visible outcome is often a one star review that mentions nobody answers the phone. Those reviews do real damage in the local results and they are almost impossible to argue with, because the complaint is true.
A reputation for being unreachable also compounds in a way rankings do not. The person who could not get through tells the neighbour. The neighbour searches, sees three reviews saying the same thing, and never calls at all. That loss never appears in any log, because the call was never placed.
Why this never comes up in a marketing meeting
The honest reason is jurisdictional. The agency is hired to generate calls and reports on calls generated. The phone is the client's problem. The answering service, if there is one, reports on calls handled and has no reason to volunteer how many it dropped. Every party is reporting a number that looks fine inside its own boundary, and the failure happens exactly at the seam between them.
So somebody has to own the whole chain, from the search result to the booked job, and look at the join nobody is paid to look at. In our own reporting we put calls answered and calls missed on the first page, above rankings, because when those two numbers are bad nothing else on the page matters.
Check it on your own site
Export one month of inbound call data
Get it from your phone provider or tracking number, not from memory. You want the timestamp, the duration and the outcome for every inbound call. If your current setup cannot produce that file, that is the first thing to replace.
Chart missed calls by hour of the day
Count anything unanswered and anything under fifteen seconds as a miss. Put them in hourly buckets. You are looking for one or two hours carrying most of the damage, which is what nearly every log shows.
Multiply out what the misses cost
Misses times your close rate times your average job value. Write the number down and keep it next to the marketing invoice, because those two figures are about to be compared.
Cover the worst hour first
One person, one shift, one specific window. Do not try to fix the whole day. Fix the hour the chart is pointing at and remeasure in two weeks so you can see whether the coverage actually changed the outcome.
Turn on a text fallback and a ten minute callback rule
The automatic text catches people who will not leave voicemail. The callback rule catches the rest, and it only works if one named person owns clearing the log before the end of the day.